Content uploaded by Muhammad Ali Al-Marhoun
Author content
All content in this area was uploaded by Muhammad Ali Al-Marhoun on Nov 19, 2015
Content may be subject to copyright.
Development of A New Correlation for Bubble Point Oil Viscosity
Abstract
An empirical correlation for oil viscosity at the bubble point was developed based on Canadian and Middle Eastern oil data. The correlation postulates a simple relationship between oil viscosity and density at the bubble point. Error analysis shows that the new correlation is superior to others when tested on the same data.
... Viscosity is an important aspect of fluid sampling as it affects the fluid flow behaviour. Viscosity is generally known as an intensive property of a fluid that causes an internal resistance of the fluid to flow 1 . Crude oil viscosity, a Newtonian fluid is a vital physical property that plays a major role in the petroleum industry, the production processing and transportation of oil due to its influence on the flow through porous rock, oil wells, multiphase flow through tubing and piping system 12 . ...
... Crude oil viscosity plays a major controlling and determining role in the successful implementation of secondary recovery process, EOR processes and reservoir simulation modeling. Optimum reservoir management and sound design facilities is hinged on reliable evaluation of viscosity data 1 . ...
Crude oil viscosity is one of the most important fluid properties that affects fluid flow behavior; either in pipeline hydraulics or in the porous media (reservoir). Viscosity is a vital physical property that plays a major role in the petroleum industry, the production processing and transportation of oil due to influence on the flow through porous rock, oil wells, multiphase flow through tubing and piping system. Therefore, the need for accurate determination of viscosity for oil and gas applications cannot be overemphasized. Numerous empirical correlations exist in literature for predicting crude oil viscosity but their accuracy is limited based on range of conditions of application, composition of the crude used in developing the correlation, specific range of data and experimental conditions. In the present work, experimental data of oil viscosity from different samples of Nigerian oil reservoirs were statistically compared with correlation predicted viscosity using the most common viscosity empirical correlations. Validity and accuracy of these empirical models has been confirmed for both saturated and under-saturated Niger Delta oil samples. It was observed that for under-saturated oil viscosities, Elshawarky & Alikhan's correlation gave a better prediction based on the Absolute average percentage error and standard deviation while for the case of saturated oil viscosities Chew and Connally proved to be the closest to the experimental results.
... Abu-Khamsin and Al-Marhoun [10] proposed a new alternative strategy to correlate bubble point oil viscosity with bubble point relative oil density. The correlation was developed using oil samples from Middle East and Canada utilizing nonlinear regression analysis. ...
Newly developed correlations for undersaturated and saturated Arabian crude oil viscosities were developed and tested using two datasets of experimental measurements. The datasets cover 71 data points of measured undersaturated viscosity (μundersaturated), pressure (P), temperature (T), bubble point pressure (Pb), gas specific gravity (γg), crude oil API, viscosity at bubble point pressure (μb) and dead oil viscosity (μd) and 79 data points of saturated viscosity (μsaturated), pressure (P), temperature (T), bubble point pressure (Pb), gas specific gravity (γg), crude oil API, viscosity at bubble point pressure (μb) and dead oil viscosity (μd), gas–oil ratio (GOR) and gas solubility (Rs). The viscosity models were developed utilizing 80% of the datasets using forward step-wise regression method. The selection of the independent variables was carried out using graphical alternating conditional expectation program (GRACE), a nonparametric regression method, which produce and generate plots for the optimal transformation of the dependent and independent variables. The program also performs low- and high-degree polynomial curve fit up to six-degree polynomial to create the desired model. The models’ accuracy was validated using the rest of the datasets, and their efficiency was tested against some commonly used correlations utilizing average absolute relative error, average relative error and cross plots. The developed models proved to be very efficient and they accurately predicted the experimental undersaturated and saturated crudes viscosities with average absolute relative errors of 1.79% and 5.89%, respectively.
... In the work of Chew and Connally [15], where they presented oil viscosity at the bubble point as a function of the solution gas-oil ratio, they used 457 data points which covered samples from South America, Canada, and the U.S. Abu-Khamsin and Al-Marhoun [18] in their work developed a correlation based on Canadian and Middle Eastern oil data, their correlation gave an average absolute error and standard deviation of 4.91% and 5.76, respectively. Also the authors Kartoatmodjo and Schmidt [19] developed oil viscosity correlation based on data bank consisting of 5321 data points while De Ghetto et al. [20] developed the saturated oil viscosity correlation based on data bank ranging from 0.07 to 295.9 cp. ...
Oil viscosity is one of the most important physical and thermodynamic property used when considering reservoir simulation, production forecasting and enhanced oil recovery. Traditional experimental procedure is expensive and time consuming while correlations are replete however they are limited in precision, hence need for a new Machine Learning (ML) models to accurately quantify oil viscosity of Niger Delta crude oil. This work presents use of ML model to predict gas-saturated and undersaturated oil viscosities. The ML used is the Support Vector Machine (SVM), it is applicable for linear and non-linear problems, the algorithm creates a hyperplane that separates data into two classes. The model was developed using data sets collected from the Niger Delta oil field. The data set was used to train, cross-validate, and test the models for reliability and accuracy. Correlation of Coefficient, Average Absolute Relative Error (AARE) and Root Mean Square Error (RMSE) were used to evaluate the developed model and compared with other correlations. Result indicated that SVM model outperformed other empirical models revealing the accuracy and advantage SVM a ML technique over expensive empirical correlations.
... Abu-Khamsim and Al-Marhoun (1991) [27] Ahmed (1989) [28] Ahmed (1992) [29] Al- Khafaji et al. (1987) [30] Al-Marhoun (1985) [31] Al-Marhoun (1988) [20] [33] Al-Marhoun (2006) [34] Al-Mehaideb (1997) [35] Al- Najjar et al. (1988) [36] Al-Shammasi (1999) [37] Andrade (1930) [38] Frick (1962) [39] Asgarpour et al. (1989) [40] Asgarpour et al. (1988) [40] Beal (1946) [41] Beggs and Robinson (1975) [42] Bennison (1998) [43] Bergman (2004) [44] Bergman (2007) [45] Bergman and Sutton (2006) [46] [47] C alhoun (1947) [48] C asey and C ronquist (1992) [49] C hew and C onnally (1959) [50] De Ghetto et al. (1995) [51] Dindoruk and C hristman (2001) [52] Doklah and Osman (1992) [53] Elam (1957) [54] El- Banbi et al. (2006) [55] Elmabrouk et al. (2010) [56] Elsharkawy and Alikhan (1997) [57] Elsharkawy and Alikhan (1999) [58] Elsharkawy and Gharbi (2001) [59] Farshad et al. (1996) [60] Fitzgerlad (1994) [61] Glasso (1980) [ Knopp and Ramsey (1960) [77] Kouzel et al. (1965) [78] Labedi (1982) [79] Lasater (1958) [82] Levitan and Murtha (1999) [83] Macary and El Batanony (1992) [84] Mazandarani and Asghari (2007) [85] McC ain (1991) [14] Movagharnejad and Fasih (1999) [91] Ng and Egbogah (1983) [92] Obomanu and Okpobiri (1987) [93] Okeke and Sylvester (2016) [94] Okoduwa and Ikiensikimama (2010) [95] Omar and Todd (1993) [24] Osorio (1990) [96] Ostermann and Owolabi (1983) [97] Owolabi (1984) [98] Oyedeko and Ulaeto (2011) [99] Petrosky (1990) [22] Petrosky and Farshad (1993) Petrosky and Farshad (1995) [101] Petrosky and Farshad (1998) [105] Standing (1947) [21] Sulaimon et al. (2014) [109] Twu (1985) [110] [111] Vasquez and Beggs (1980) [112] Velarde et al. (1997) [113] Whitson and Brule (2000) [4] ...
11 Precise PVT studies and performance of phase-equilibria for petroleum reservoir fluid are essential for 12 describing these fluids and appraising their volumetric behavior at several pressure stages. There are 13 numerous laboratory studies that can be generated in a reservoir sample. The amount of available 14 data regulates the number of tests that can be achieved in the laboratory. Generally, there are three 15 laboratory tests that characterize hydrocarbon fluids, including primary study, constant mass 16 depletion, and differential vaporization test. Generally, PVT properties determined either 17 experimentally or calculated theoretically through published correlations. In this chapter, the author 18 details different PVT laboratory tests utilized to distinguish the phase b ehavior of black oil. 19
Challenges during CO2 flooding of oil reservoirs are mobility control and gravity override caused, respectively, by lower viscosity and density of CO2 compared with reservoir fluids. Blocking agents such as foam, cross-linked polymers and gels have been tested. However, such techniques are more effective in preventing fingering and blocking thief zones than reducing gravity override. Moreover, no foaming agent has been identified to withstand the high temperatures and salinities of some reservoirs. This experimental study investigates the use of triethyl citrate (TEC) as a densifying agent to blend with supercritical CO2 (scCO2) to minimize gravity override. High-permeability, Indiana limestone core samples saturated with either dead or live light crude oils at irreducible water saturations were flooded vertically upward at 24.132 MPa and 100 °C with about one pore volume of either pure scCO2 or a scCO2–TEC blend of 0.15 mol fraction (0.525 mass fraction) TEC. At breakthrough, scCO2 flooding of dead oil recovered 14.82% of the initial oil in place, while the blend flood recovered 31.99%. At 1 PV, the scCO2 flood recovered 60.14%, while the blend flood recovered 72.67%. The live-oil-saturated core samples produced similar results but with generally lower oil recoveries. At 1 PV of fluid injected, all floods reduced the oil saturation to levels lower than the residual oil saturation achieved with water flooding. With TEC’s low solubility in both crude oil and water, blending scCO2 with TEC is, thus, a technically viable method to mitigate gravity override in CO2–EOR process.
Accurate knowledge of reservoir fluid properties, especially reservoir saturation pressures and reservoir type, is paramount for the estimation of reservoir volumetrics, well design and placement, well and reservoir performance management, field development planning and ultimately economic evaluation of the reservoir. However, most of the existing methods such as PVT fluid sampling, compositional grading, and empirical models have proven to be either ineffective or expensive and sometimes, leads to ungeneralizable results.
This paper discusses the application of machine learning (ML) techniques to develop a robust model for prediction of reservoir fluid properties such as saturation pressures in a Niger Delta Field and the subsequent classification of the reservoirs as under-saturated or saturated. Reservoir data including the PVT data, compositions (C1-C7+), temperature & pressure data, fluids contacts of known reservoir type were considered initially in order to train a model using a multi-features regression ML algorithms for the determination of the saturation pressure and a Two-class boosted decision ML algorithms to determine the determine the type of reservoir (saturated or undersaturated). Of the 34 parameters considered, it was found that the reservoir fluid composition has significant impact in determining the accuracy of the Pb-hypothesis. The key performance indicators such as MAE, RMSE, and RSE are within 0.02-0.05 and Co-efficient of determination of about 95% for Pb determination. When compared with the Standing, Glaso, Petrosky-Farshad etc. correlation, the AAE was significantly less than both cases. AAE for the Standing and Glaso correlations were respectively18.5% and 25.1% while that of the ML model was 2.3% using both the data from training set and test set. For the classification algorithms in determining type of reservoir, the model performed within the 73-100% accuracy, precision & recall. The Area under the curve (AUC) of the Receiver Operator Characteristic (ROC) chart of approximately 97% indicated the robustness of the model. The results showed that the use of a properly trained and accurately validated ML model can deliver better predictions of reservoir fluid properties and subsequent reservoir type when compared to conventional methods.
With the cusp in the advances in Artificial Intelligence and Machine Learning, there have been several development of AI-based PVT models. Although they have offered some advantages in one way or the other, their easy implementation and flexibility have been limited due to computational complexities, and the relatively vast amount of memory allocation required. Also, some of the classical PVT models are wrought with certain setbacks. Some of these models are only accurate within specific conditions of crude-oil type and properties, and hence fail when there are small variations in crude oil properties or when there are different genetic sources or prevalence for the crude oil type. Here in this paper, we have developed a novel algorithmic design and implementation based on a combination of mathematical combinatorics and linear algebra. Using the superposition principle in an ingenious way, we built new, improved and better models from existing PVT correlations and models. Our resulting algorithm known as IntelliPVT offers more robust, more accurate and faster PVT model predictions. IntelliPVT has been written both in MATLAB and Python. Using real field data published in literatures, journals and books, and compared with other realistic conventional PVT correlations, IntelliPVT outperformed these correlations and predicted Crude-Oil PVT properties of Formation Volume Factors, Solution Gas-Oil Ratios, Bubble Point Pressures, Viscosities, and Crude oil compressibilities to the least Root Mean Square Error, R-Squared values between 0.91 and 0.99, and the lowest minimum mean square error with very high computational speed. The novel algorithm offers more accurate results with a faster implementation and applicability to all crude-oil types including heavy oils. Its implementation will ensure a better subsurface characterization, improved reservoir management as well as surface design along the whole spectrum of the oilfield lifecycle.
Precise PVT studies and behavior of phase-equilibrium of petroleum reservoir fluids are essential for describing these fluids and appraising their volumetric behavior at several pressure stages. There are numerous laboratory studies that can be performed on a reservoir sample. The amount of data desired determines the number of tests to be performed in the laboratory. Generally, there are three laboratory tests which characterize hydrocarbon fluids, namely primary study, constant mass depletion, and differential vaporization test. Generally, PVT properties are determined either experimentally or calculated theoretically through published correlations. This chapter presents different PVT laboratory tests that are required to understand the phase behavior of black oils.
Density and viscosity are of the most important governing parameters of the fluid flow, either in the porous media or in pipelines. Ideally, viscosity and density determined experimentally in the laboratory on actual fluid samples taken from the field under study. However, in the absence of experimentally measured data, especially during the prospecting phase, or when only invalid samples are available, one can resort to empirically derived PVT correlations. So it is of great importance to use accurate correlations to calculate the crude oil density and viscosity at various operating conditions. During the last decades, several correlations have been developed to estimate density and viscosity of oil at different reservoir conditions. However, these correlations may be useful only in regional geological provinces and may not provide satisfactory results when applied to crude oils from other regions since oil properties differ according to its source , origin and core type .Also, crude oil composition is complex and often undefined. Therefore, based on Egyptian oil reservoirs data; new correlations have been developed for predicting density and viscosity of dead and live crude oil. Validity and accuracy of these correlations have been confirmed by comparing the obtained results of these correlations and other ones with experimental data for Egyptian oil samples. Checking results of these correlations show that correlations developed by this study revealed more accurate results than the literature correlations.
ResearchGate has not been able to resolve any references for this publication.